Skip to content

New mechanism for oxidation of native silicon oxide

Evidence and attribution

Authority of statements

Prose below summarizes the publication identified by doi, title, and pdf_path.

Summary

Reactive atomistic simulations study hyperthermal oxygen (1–5 eV) interacting with thin native SiO\(_x\) (\(x \le 2\)) films (~10 Å) on Si(100)–(2×1). Room temperature: penetrant oxygen resides mainly inside the oxide, not at the SiO\(_x\)|c-Si interface. ≥ ~700 K: oxygen diffuses through the oxide and reacts at the crystalline Si boundary—closer to thermal oxidation kinetics described by Deal–Grove at high \(T\). The article also analyzes defect formation during oxidation relevant to ultrathin gate dielectrics and discusses how oxidant energy and temperature steer mechanism. Readers should cross-check any quantitative barrier or rate statements against the figures and tables in the primary PDF rather than relying on this synopsis alone.

Methods

1 — MD application (atomistic dynamics). Molecular dynamics simulations use a Si|SiO\(_x\) ReaxFF parameterization (Buehler-type Si oxidation force field, as cited in the article). A native oxide slab stack on Si(100)-(2×1) is described with an initial cell about 21.7 × 21.7 × 27.1 Å containing Si and O atoms arranged as a thin SiO\(_x\) film, prepared by oxidizing with 1 eV O\(_2\), then equilibrated at 300, 700, and 900 K under NVT with a Berendsen thermostat (20 ps, damping 0.1 ps) followed by 10 ps NVE; the authors report no significant difference between Berendsen and Bussi thermostats for this equilibration test. Hyperthermal O/O\(_2\) with 1–5 eV beam conditions impinges on the thin SiO\(_x\) film to probe localization vs interface reaction vs temperature (papers/Khalilov_Si_oxidation_JPCC_2013.pdf; normalized/extracts/2013u-khalilov-j-phys-chem-new-mechanisms_p1-2.txt). Engine / code: N/A — MD software not named in the indexed excerpt (confirm LAMMPS or other engine in pdf_path). Timestep: N/A — not in the indexed excerpt. Boundaries / PBC: N/A — not spelled out in the excerpt used here. Barostat: N/A — NVT/NVE legs described without hydrostatic NPT control in this summary layer. Pressure targets: N/A — no GPa/bar hydrostatic protocol in the excerpted equilibration description. Electric field: N/A — not used in the excerpted protocol description. Replica / enhanced sampling: N/A — not used.

2 — Force-field training. N/A — the page applies a literature ReaxFF parametrization for Si/O oxidation rather than reporting a new training workflow in the excerpted material.

3 — Static QM / DFT-only. N/A — reactive MD drives the hyperthermal oxidation study; DFT may appear elsewhere in the article for benchmarks, but is not summarized as the production engine here.

Findings

Outcomes & mechanisms. At room temperature, penetrant oxygen resides mainly inside the oxide, not at the SiO\(_x\)|c-Si interface. At ≥ ~700 K, oxygen diffuses through the oxide and reacts at the crystalline Si boundary—closer to thermal oxidation kinetics discussed with Deal–Grove-type expectations at high \(T\). Oxidant energy and temperature steer defect formation and interface chemistry relevant to ultrathin gate stacks.

Comparisons. The synopsis ties hyperthermal localization behavior to high-temperature diffusion-limited oxidation pictures; detailed experimental comparison belongs to the PDF Discussion.

Sensitivity & design levers. Substrate temperature (300 / 700 / 900 K equilibration window cited above) and hyperthermal 1–5 eV oxidant conditions control where oxygen localizes vs reacts at the Si interface.

Limitations & outlook. Model-size and classical-reactive approximations limit quantitative transfer to every experimental beam system; the article spans distinct hyperthermal vs thermal regimes—readers should not merge them without the authors’ caveats in the PDF.

Corpus honesty. Integration timestep and full PBC specification are not in the _p1-2 extract; confirm in pdf_path for reproduction-grade detail.

Limitations

Model sizes and classical reactive approximations limit direct quantitative agreement with every experimental beam system. The study spans hyperthermal impingement and elevated temperature diffusion regimes; operators should not collapse those into a single “oxidation mechanism” without reading the temperature-dependent localization discussion in the article.

Relevance to group

Silicon oxidation with van Duin ReaxFF in the PLASMANT collaboration—bridging hyperthermal and thermal oxidation pictures. Connects to gate dielectric processing discussions where beam energy and substrate temperature jointly determine whether oxidation localizes in the oxide or reacts at the Si interface.

Citations and evidence anchors

  • DOI: 10.1021/jp400433u
  • Extract: normalized/extracts/2013u-khalilov-j-phys-chem-new-mechanisms_p1-2.txt